[dns-operations] DNS software test!

[延志伟]

Hi, Phil

Thank you very much for your reply.

In my test, the 50000 query entries all belong to the second level domain. However, the TTL in root, top level domain (test.) and second level domain (my.test.) all set to 0 to prevent the caching.

When the NSD is deployed, I think the case2 and case3 should have similar results. However, they are definitely different and that is my confusion.

Case2  Recursive (bind)--Root (bind)--TA (bind)-- SA (nsd)

173.912569 qps

Case3  Recursive (bind)--Root (bind)--TA (nsd)-- SA (bind)

1857.768915 qps

Zhiwei








在 2011-09-16 14:17:35，"Phil Regnauld" <regnauld at nsrc.org> 写道：
>延志伟 (yzw_iplab) writes:
>> Hi, all,
>> 
>> Client executes the DNS query using ‘queryperf’ under the following four scenarios (where SA and TA denote the Second levelAuthority server andTop levelAuthority server respectively)，their QPS results are：
>> 
>> Case1：Recursive(bind)àRoot(bind)àTA(bind)àSA(bind)
>> 
>> As you find, only when the NSD is deployed on the TA, the performance (e.g., QPS) can be promoted significantly. In other words, even when the SA uses the NSD, the QPS is just similar with the case when all the servers use BIND.
>
>	I'm not sure queryperf is meant to be tested in this fashion - you
>	have too many factors interacting for one thing, and we don't really
>	know the configuration of each of the servers (for example, have you
>	turned on early prefetch on about to expire records on unbound, what's
>	the query size, what data set are you querying against - live or synthetic
>	data, etc..)
>
>	Testing one tool at a time, on a known dataset would be the way to
>	go for recursion (unbound, bind) and authoritative (bind, nsd).
>
>	Cheers,
>	Phil






-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.dns-oarc.net/pipermail/dns-operations/attachments/20110916/47b17706/attachment.html>